TYPE Original Research PUBLISHED 20 October 2022 DOI 10.3389/fvets.2022.980414 Clinical and magnetic resonance OPEN ACCESS imaging features, and pathological findings of spinal EDITED BY Lisa Alves, University of Cambridge, United Kingdom REVIEWED BY lymphoma in 27 cats Elsa Beltran, Royal Veterinary College (RVC), United Kingdom Valentina Lorenzo1*, João Ribeiro2,3 , Marco Bernardini4,5 , Marcelo Gomez, Austral University of Chile, Chile Juan J. Mínguez6,7 , Meritxell Moral1 , Carlos Blanco1 , *CORRESPONDENCE Tina Loncarica4 , Araceli Gamito6 and Martí Pumarola8 Valentina Lorenzo 1 valo.lorenzo@gmail.com Neurología Veterinaria, Getafe, Madrid, Spain, 2 Referência Veterinária, Cascais, Portugal, 3 Faculdade de Medicina Veterinária da Universidade Lusófona de Humanidades e Tecnologias de SPECIALTY SECTION Lisboa, Lisbon, Portugal, 4 Anicura Ospedale Veterinario I Portoni Rossi, Zola Predosa, Italy, This article was submitted to 5 Department of Animal Medicine, Production and Health, University of Padua, Padua, Italy, 6 Hospital Veterinary Neurology and Veterinario Guadiamar, Sevilla, Spain, 7 Scarsdale Vets-Pride Veterinary Centre, Derby, Neurosurgery, United Kingdom, 8 Mouse and Comparative Pathology Unit, Department of Animal Medicine and a section of the journal Surgery, Veterinary Faculty, Campus Universitat Autónoma de Barcelona (UAB), Barcelona, Spain Frontiers in Veterinary Science RECEIVED 28 June 2022 ACCEPTED 27 September 2022 This multicentric retrospective study describes the clinical and MRI features PUBLISHED 20 October 2022 and pathological studies of spinal lymphoma in 27 cats. MRI characteristics CITATION Lorenzo V, Ribeiro J, Bernardini M, and their possible correlations with histopathological findings were studied. Mínguez JJ, Moral M, Blanco C, The most frequent neurological signs were rapidly progressive paraparesis Loncarica T, Gamito A and (62.9%) or paraplegia (22.2%). Bimodal age distribution was found with 40.7% Pumarola M (2022) Clinical and magnetic resonance imaging features, of cats aged ≤2.5 years (63.6% of them FeLV positive), and 44.4% of cats and pathological findings of spinal aged ≥8 years (16.7% of them FeLV positive). Spinal lymphoma was generally lymphoma in 27 cats. Front. Vet. Sci. 9:980414. presented on MRI as an ill-defined epidural focal lesion with moderate to doi: 10.3389/fvets.2022.980414 severe spinal cord compression, expanding more than one vertebral body. COPYRIGHT MRI lesions were typically localized in the lumbar vertebral segment (p = © 2022 Lorenzo, Ribeiro, Bernardini, 0.01), circumferential to the spinal cord (p = 0.04), hyperintense on T2- Mínguez, Moral, Blanco, Loncarica, Gamito and Pumarola. This is an weighted sequences (p = 4.3e-06), and isointense on T1-weighted sequences open-access article distributed under (p = 8.9e-07). The degree and pattern of contrast enhancement were variable. the terms of the Creative Commons Other morphological patterns included paravertebral masses with extension Attribution License (CC BY). The use, distribution or reproduction in other into the vertebral canal and lesions centered in the spinal nerve roots. forums is permitted, provided the Involvement of vertebrae and adjacent spinal soft tissues was present in 74% original author(s) and the copyright owner(s) are credited and that the of cases when present vertebral involvement was characterized by cortical original publication in this journal is sparing. When follow-up MRI studies (n = 4) were performed after treatment cited, in accordance with accepted new lesions of similar nature but different localizations and extension were academic practice. No use, distribution or reproduction is permitted which observed. Confirmation of spinal lymphoma was performed by CSF analysis in does not comply with these terms. 4/27 (14.8%) of cases, by FNA in 6/27 (22.2%) of cases, by surgical biopsy in 10/27 (37%) of cases, by FNA and surgical biopsy in 1/27 (3.7%) of cases, by CSF, FNA, surgical biopsy and postmorten examination in 1/27 (3.7%) of cases, and postmorten studies in 5/27 (18.5%) of cases. Antemortem diagnosis was achieved in 22/27 (81.5%) cats. The presence of necrosis in histopathological studies as an unfavorable prognostic indicator of survival was significantly more probable when lesions were not hyperintense on T2-weighted sequences (p = 0.017). Spinal lymphoma in cats is a complex entity with heterogeneous Frontiers in Veterinary Science 01 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 imaging and histopathological appearance. However, certain MRI features may support a tentative diagnosis, which in a group of cases can be confirmed when combined with the CSF findings. For the rest of the cases, tissue sampling assisted by imaging findings remains necessary for definitive diagnosis. KEYWORDS feline, MRI, imaging, neoplasia, spinal cord, CSF, cytology, histopathology Introduction MRI was confirmed based on evidence of neoplastic cells in the CSF and/or in tumor samples. Cases were not included if the Lymphoma is the most common spinal neoplasia in cats results were inconclusive. (1–3) and the second most common disease of the spinal cord Information regarding age, sex, feline leukemia virus (1, 2, 4, 5) but information about Magnetic Resonance Imaging (FeLV) status, neurological evaluation, duration of clinical (MRI) features is still scarce. To date, case series of feline spinal signs prior to MRI studies, survival time, treatment [medical lymphoma have been limited to a few reports including limited management -steroids, non-steroidal anti-inflammatory drugs cases (3, 6–8). (NSAIDs), antibiotics, analgesic, or some combination of these-, The few descriptions of MRI findings, together with the chemotherapy, surgery, or combination of these] and outcome heterogeneous histopathological nature of this disease (9) (survival time from the day of MRI) was retrieved. hampers precise prognosis and clinical management. MRI, as Cases were categorized by tumor location in relation to being the gold standard technique for assessing the nervous the vertebral column on MRI, and together with the clinical system, is highly valuable for prompt diagnosis of neoplasia. data, a presumptive diagnosis was generated. Cases were finally In fact, in humans, presumptive antemortem diagnosis of diagnosed and further subclassified into tumor types according lymphoma affecting the nervous system is based on the to clinicopathological and anatomopathological studies. identification of MRI-specific morphologic features (10–12), The sampling and the imaging procedures were held in and the correlation of MRI features in primary brain lymphoma a clinical environment as part of the diagnostic protocol for with histopathologic findings has been studied (12). Moreover, medical purposes, with the signed consent of the owners. The as MRI allows the characterization of the distribution of material used for the study did not require supplementary the lesions it may facilitate tissue sampling for diagnosis procedures or handling. confirmation by cytology or histopathology and therefore facilitate adequate treatment and prognosis evaluation. The goals of this study were to describe the clinical and MRI study acquisition and evaluation MRI characteristics of feline spinal lymphoma in 27 cats and to compare the findings with previous reports in an MRI studies were performed using 3 different 1.5-Tesla attempt to identify characteristics that may aid in antemortem magnets (Gyroscan Intera and Prodiva, Philips, Eindhoven, The diagnosis. Moreover, possible correlations of MRI features Netherlands; Vantage Elan, Canon Medical Systems Europe, with the histopathological findings which may assist prognosis Zoetermeer, The Netherlands), a 0.5-Tesla magnet (Gyroscan, were investigated. T5-II, Philips, Eindhoven, The Netherlands) and 3 different 0.2- Tesla magnets (MrJ, Paramed, Genoa, Italy; Hitachi Airis mate, Chiyoda, Japan; Vet MRI, Esaote, Genoa, Italy). Materials and methods All MRI studies were performed under inhalation anesthesia with the patient in dorsal recumbency. The area of study Medical records including the clinical data, the MRI studies of the vertebral column was based on the clinical findings with their corresponding reports, and the results from the and neurolocalization and extended if necessary according pathology studies of spinal lymphoma affecting the nervous to the findings on each particular case. MRI protocols system in cats from 5 neurology referral centers were retrieved included at least T2-weighted (T2W) sequences in sagittal and by a board-certified or a residency-trained neurologist. MRI transverse planes and T1-weighted (T1W) sequences before and studies were reviewed by two authors blinded to the final after intravenous administration of gadolinium-based contrast diagnosis at the time of image evaluation and the final result was medium (Gadoteric acid, Dotarem© , Guerbet, or Claricyclic© , achieved by consensus. All the information was retrospectively GE Healthcare, at a dose of 0.1 mmol/Kg). reviewed by a board-certified neurologist (VL). Cases were Cases were categorized by tumor location in reference included in the study if presumptive diagnosis on spinal cord to the spinal cord (intramedullary, intradural extramedullary, Frontiers in Veterinary Science 02 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 epidural, spinal nerve root/s) and located in relation to the cells were considered in order to avoid possible diagnostic corresponding vertebral body/ies. Furthermore, lesions were errors which could be induced by the morphological overlapping classified as solitary (focal or extensive) or multifocal. The lesion between small neoplastic lymphocytes and reactive lymphocytes was considered extensive if it extended over the length of >1 (8, 14). vertebral body considering the vertebral canal component. When possible, further classification was done based The signal intensity on T1W, T2W, and additional STIR on the Revised European–American Lymphoma (REAL) sequences was defined as iso- hypo- or hyperintense compared classification of lymphoid neoplasms adopted by the World to normal spinal cord parenchyma. Health Organization (WHO) (15, 16). According to cell The degree of contrast enhancement when pre- and post- morphology large, intermediate or small cell tumor types were contrast images were compared was subjectively defined as none, established and used to classify the tumor types in grades mild, moderate, or intense, and the type as homogeneous or according to the National Cancer Institute working formulation heterogeneous. Meningeal contrast enhancement was recorded (NCI WF) (16). The most recently published data on animal if present. nervous system tumors was also considered (17). Additional recorded features included the presence and On histology, the data recorded included: morphology of the degree of spinal cord compression (mild, <25%; moderate, 25– cells, mitotic count/40X (low 0–1, moderate 2–4, and high ≥5), 50%; severe, >50% occupation of the vertebral canal diameter) presence of degenerative changes (necrosis and/or hemorrhage), and other possible spinal cord changes. and tissue infiltration. Results of immunohistochemistry (IHQ) When the vertebral bone or other adjacent bone structure were also recorded and, when performed, tumors were classified was involved, changes in the bone marrow pre and post- as B-cell or T-cell lymphoma according to immunoreactivity contrast images, the presence of cortical lysis (discontinuity of of the cells in front of CD20 and CD3 markers, respectively, the normal cortical margin), cortical sparing (preservation of following published protocols (18). the normal cortical margin), and preservation of vertebral shape Lymphomas were categorized into low, intermediate, or (preservation of the overall architecture of the vertebra despite high grades (16). Small cell lymphoma cells had nuclei 1–1.5 small foci of cortical lysis) were studied. If present, involvement times the diameter of the red blood cells whereas in large cell of soft tissues (paraspinal and/or distant) was registered. lymphoma nuclei cells were 2–3 times the diameter of the red Follow-up MRI studies after surgery and/or chemotherapy blood cells. The small-cell lymphomas, including the small were performed on 4 cats. When a second MRI was performed lymphocytic lymphoma, small-cleaved cell lymphoma, small findings on the area of the previous lesion/s and new findings lymphocytic lymphoma with plasmacytoid differentiation, were recorded. and the small lymphocytic lymphoma with intermediate differentiation, form the low-grade lymphoid neoplasms. The large-cell and large cleaved-cell lymphomas form the CSF and tissue studies intermediate-grade grouping. The high-grade lymphomas include lymphoblastic and small non-cleaved-cell lymphomas. Diagnosis was confirmed either by cytology (from CSF or When complete necropsy examination was available the lesions), histopathology (from a surgical biopsy or necropsy), or presence of lesions involving other organs (multicentric by a combination of them. The tissues were sampled for cytology lymphoma) or not (primary lymphoma) was recorded. by fine needle aspirate (FNA) and/or imprint. Data regarding the method of diagnosis and results were recorded. Diagnosis of lymphoma was based on the observation of a population of Statistical analysis morphologically abnormal lymphoid cells. The CSF analysis was considered abnormal if the results The characteristics from the MRI studies were retrieved were above the reference values of nucleated cell count to obtain explanatory variables. Data were arranged into (NCC) ≤5 cells/µL, total protein (TP) concentration for contingency tables and the frequency distribution of the cisternal puncture ≤25 mg/dL and lumbar puncture ≤45 variables was assessed by a chi-square goodness of fit test mg/dL, respectively, and/or abnormal cell morphology or using a R software. Results were considered significant if p- distribution. An increase in TP with normal NCC was defined value was <0.05. In order to correlate the MRI features to the as albuminocytological dissociation. On cytology, a diagnosis of histological findings, the presence of necrosis was chosen as the lymphoma was considered if a population of lymphoid cells with dependent variable as this parameter was previously used for this atypical morphology was observed, which can include increased purpose (12), and it can be related to the tumor grading, with cell size (>2.5 times the upper limit of normal), irregular shape, higher grade lymphomas generally showing a higher incidence and pointed borders of the cytoplasm, and nuclear atypia with of necrosis (16). The explanatory variables were converted to any degree of sharp nuclear notches (13). When the diagnosis binary and a Stata probit model was run to study the correlation. was based on cytology, only the samples with atypical large Results were considered significant if p-value was <0.05. Frontiers in Veterinary Science 03 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 TABLE 1 Signalment and clinical data of 27 cats with spinal lymphoma. Breed Sex Age, FeLV test Duration of Initial signs Clinical signs at median (n = 21) signs presentation (range) (n = 21) DSH 24/27 Male 13/27 5 years (5 Positive 9/21 ≤7 days 12/21 Lameness 1/27 Spinal pain 9/27 Maine Coon 2/27 Female months−15 Negative 12/21 8 days-1 month Progressive Neurological sings 27/27: Turkish Van 1/27 14/27 years) 7/21 paresis 26/27 – ambulatory paraparesis 10/27 1–3 months 2/21 – non-ambulatory paraparesis 7/27 – paraplegia 6/27 – tetraparesis 2/27 – thoracic limb monoparesis 1/27 – tetraplegia 1/27 Results MRI features Study population and clinical data MRI studies were performed using high-field 1.5-Tesla magnets in 11 cats, a medium-field 0.5-Tesla magnet in 2 cats, A total of 27 cats were included in the study. Results are and low-field 0.2-Tesla magnets in 14 cats. summarized in Table 1. The localization and characteristics of the lesions are Breeds included domestic shorthair (n = 24), Maine Coon summarized in Tables 2, 3. (n = 2), and Turkish Van (n = 1). Age ranged between 5 months On the first presentation, an epidural solitary lesion was and 15 years with a mean age of 5.9 years and a median of 5 present in 12/27 (44.4%) cats. The lesion was localized in years. In this group of cats, 11/27 (40.7%) had ages comprised the cervical vertebral column in 1/12 cases, the thoracic between 5 months and 2.5 years, and 12/27 (44.4%) comprised vertebral column in 5/12 cases, and the lumbar vertebral between 8 and 15 years, the remaining 4 cats were aged 4 column in 6/12. In 3 of these 12 cases, a second MRI (2 cases), 5 and 6 years. Fourteen cats were females and 13 (2 after surgery and 1 after surgery and chemotherapy) cats were males. FeLV test results were recorded for 21 cats, was performed due to recurrence of signs, and new lesions results were positive for 9/21 cats (42.9%) and negative for with differences in location and extension were found. 12/21 (57.1%). Results are summarized in Table 3. In 1/27 (3.7%) cases The time-lapse between the onset of the clinical signs and multifocal epidural lesions were present (at T13-L1, L4-L5, and the MRI study was retrieved for 21 cats. For 12/21 (57.1%) L5-L6 vertebrae). the duration was ≤7 days, for 7/21 (33.3%) ranged from 1 Two separate lesions, one epidural extending from the week to 1 month, and for 2/21 (9.5%) of cases the duration conus medullaris to the first coccygeal vertebrae and another of the signs was 3 months. The mean was 17.6 days (range intramedullary lesion extending from the L2 vertebra to the 1–90 days). conus medullaris were found in 1/27 (3.7%) cases (Figures 1A– According to the anamnesis, the owner’s initial presenting H). The cat was treated with chemotherapy (COP protocol) complaint was signs of thoracic or pelvic limb lameness for with improvement, in a control MRI that was done 3 1/27 (3.7%) cats and progressive paresis for 26/27 (96.3%) months later a residual intramedullary malacic lesion on cases (paraparesis 23/27, 85.2%, tetraparesis 2/27, 7.4% and the L5 vertebra was the only relevant finding (Figures 1I– monoparesis 1/27, 3.7%). The clinical signs on evaluation P) (case 4 in Table 3). At 9 months, an MRI of the head at the referral centers included spinal pain in 9/27 (33.3%) was done due to recent diffuse intracranial signs, showing a cats, ambulatory paraparesis in 10/27 (37%), non-ambulatory diffuse and intense meningeal enhancement compatible with paraparesis in 7/27 (25.9%), paraplegia in 6/27 (22.2%), lymphomatous meningitis. Further studies were not allowed by tetraparesis in 2/27 (7.4%), thoracic limb monoparesis in 1/27 the owners. (3.7%), and tetraplegia in 1/27 (3.7%) of cases. A single lesion with intramedullary/intradural/epidural Previous use of corticosteroids (prednisone or prednisolone) involvement (located at L5-L6 vertebrae) was present in 1/27 was recorded for 12 cats and of NSAIDs for 8 before diagnosis. (3.7%) cases and two separate intramedullary lesions (located In all the cases a temporary partial improvement was observed at T2-T4 vertebrae and T9-T11 vertebrae) were present in 1/27 except in one cat from the NSAIDs group. (3.7%) cat. Frontiers in Veterinary Science 04 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 TABLE 2 MRI distribution and pattern of spinal lymphoma in 27 cats on presentation. Number Localization Extense Spinal cord T2W T1W Contrast Enhancement (%) (+1 vert) compression enhancement pattern Epidural focal 12/27 (44.4) Thoracic 5/12 8/12 Moderate 7/12 Hyper 8/12 Iso 7/12 Intense 6/12 Homog 7/11 Lumbar 6/12 Severe 5/12 Iso 3/12 Hyper 4/12 Moderate 2/12 Heterog 4/11 Cervical 1/12 Iso-hyper 1/12 Iso-hyper 1/12 Mild 3/12 None 1/12 Epidural 1/27 (3.7) Lumbar 3/3 1/3 Mild 2/3 Hyper 3/3 Iso 3/3 Mild 3/3 Homog 3/3 multifocal (3 Moderate 1/3 lesions) Epidural and 1/27 (3.7) Epidural: 2/2 Mild 1/1 Hyper 2/2 Iso 2/2 Intense 2/2 Homog 1/2 intramedullary connus-Cc1 (epidural) multifocal (2 intramedullary: Heterog 1/2 lesions) L2-conus (intramed.) Epidural/intradural/ 1/27 (3.7) Lumbar 1/1 0/1 Mild 1/1 Hyper 1/1 Iso 1/2 Intense 1/1 Heterog 1/1 in tramedullary Intramedullary 1/27 (3.7) Thoracic 2/2 2/2 - Hyper 2/2 Iso 2/2 Mild 1/2 Intense homog 2/2 multifocal (2 1/2 lesions) Nerve roots 5/27 (18.5) Cervical 2/5 4/5 Moderate 3/5 Iso-hyper 1/5 Iso 3/5 Intense 4/5 Homog 3/5 intracanal Lumbar 2/5 Severe 2/5 Hyper 4/5 Iso-hyper 2/5 Moderate 1/5 Heterog 2/5 lumbar-S3 1/5 Intramedullary 1/27 (3.7) Cervical 2/2 0/2 Mild 1/1 Hyper 2/2 Iso-hyper 2/2 Mild 2/2 Heterog 2/2 and nerve roots (2 lesions) Paravertebral 4/27 (14.8) Thoracic 4/4 3/4 Moderate 1/4 Iso 1/4 Iso 4/4 Mild 1/4 Homog 2/4 mass and epidural Severe 3/4 Iso-hyper 1/4 Moderate 2/4 Heterog 2/4 focal Hyper 2/4 Intense 1/4 Epidural and 1/27 (3.7) Epidural: L5-S1 2/2 Moderate 1/2 Hyper 2/2 Iso 2/2 Mild 2/2 Heterog 2/2 paravertebral Paravertebral Severe 1/2 mass multifocal T12-L3 (2 lesions) Total Focal 22/27 Cervical 5/33 22/33 Mild 5/29 Hyper 26/33 Iso 24/33 None 1/33 Homog 18/32 (81.5) (15.15) Moderate 13/29 Iso 4/33 Iso-Hyper 5/33 Mild 12/33 Heterog 14/32 Multifocal Thoracic 12/33 Severe 11/29 Iso-hyper 3/33 Hyper 4/33 Moderate 5/33 5/27 (18.5) (36.36) Intense 15/33 Lumbar- lumbosacral 16/33 (48.48) Extraneural involvement – Bone 7/27 (26) – Soft tissue 11/27 (40.7) Hyper, hyperintense; Iso, isointense; Homog, homogeneous; Heterog, heterogeneous. Lesions involving nerve roots and causing spinal cord lesion located at C4-C5 and a separate lesion involving the right compression were present in 5/27 (18.5%) cases, localized in C6-C7-C8 nerve roots was observed in 1/27 (3.7%) cases. the right C5 to C8 nerve roots, the right C6-C7 nerve roots A paravertebral mass with vertebral canal invasion along (Figure 2), bilateral L5-L7 nerve roots 1/5, right L6 and left L7, nerve roots was found in 4/27 (14.8%) cases, localized at T2- and bilateral L5-S3 (Figure 3), respectively. An intramedullary T4 vertebrae, T5-T7, T8-T9, and T6-T12 vertebrae, respectively. Frontiers in Veterinary Science 05 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 TABLE 3 MRI distribution and imaging pattern of spinal lymphoma on follow-up MRI performed in 4 cats. Localization Treatment Time Extense Spinal cord T2W T1W Contrast Extraneural between (+1 vert) compression enhancement findings MRI Case 1 Epidural half L2 Surgery 18 days Yes Moderate Iso Iso Intense homog. Bilateral 1st MRI to L3-L4 sublumbar lateralized to right muscle enhancement 2nd MRI Epidural half L2 Yes Severe Iso-hyper Iso Intense homog. to half L4 lateralized to left Case 2 Epidural L4-L5 Surgery 13 months No Severe Iso-hyper Hyper Intense heterog. None 1st MRI lateralized to the left 2nd MRI Paravertebral No Mild Iso-hyper Hyper Intense heterog. Right iliac crest, lumbosacral right, L6 and L7 L7 nerve root vertebrae Case 3 Epidural T8-T10 Surgery and 3 months Yes Moderate Iso Iso None None 1st MRI left chemotherapy 2nd MRI Epidural 2 Yes Mild Iso Iso Mild homog. none lesions: T8 to T9 No Mild Iso Iso Mild homog right and L3 right Case 4 Epidural Chemotherapy 3 months Mild Hyper Iso Intense epidural L7 vertebral 1st MRI connus-Cc1 and homog. intramedullary Intramedullary L2-conus heterog. 2nd MRI Malacic lesion L5 None None None Hyper, hyperintense; Iso, isointense; Homog, homogeneous; Heterog, heterogeneous. In 1/27 (3.7%) case a paravertebral mass with vertebral canal associated with the extramedullary lesions was mild in 8/33 invasion along nerve roots in T12-L3 vertebrae and a separate (24.24%), moderate in 13/33 (39.39%), and severe in 12/33 epidural lesion in L5-S1 vertebrae were depicted (Figure 4). (36.36%) cases. In 22/27 (81.5%) cases a focal lesion was found, and 5/27 On relaxometry, lesions were hyperintense (26/37, 70.27%) (18.5%) cases had multifocal lesions. Considering the 3 cases isointense (6/37, 16.21%), or iso-hyperintense (5/37, 13.51%) with recurrence in the follow-up studies lesions were focal in relative to the spinal cord on T2W sequences. Isointensity 24/30 (80%) and multifocal in 6/30 (20%). on T2W images was seen in cases with solitary epidural A total number of 33 lesions were observed in the 27 patients lesions in 5/6 (Figures 5A, 6A) and 1/6 cases with paravertebral on presentation, and 4 new lesions were observed in the follow- mass with epidural extension. The 5 iso-hyperintense lesions up MRI studies thus making a total of 37 lesions. Lesions were corresponded to 2/5 epidural, 1/5 nerve roots, and 2/5 present at the level of the lumbar and lumbosacral (19/37, paravertebral mass. A case with an epidural isointense lesion 51.35%), thoracic (13/37, 35.13%), and cervical (5/37, 13.51%) had an increased signal with iso-hyperintensity on a follow- vertebrae. Extensive lesions (larger than 1 vertebral body) were up MRI (case 1 in Table 3; Figure 6). On T1W sequences present in 24/37 (64.86%) cases. Lesions were circumferential lesions were isointense (27/37, 72.97%), iso-hyperintense (5/37 to the spinal cord in 13/37 (48.15%), lateralized to the right 13.51%) or hyperintense (5/37, 13.51%). Iso-hyperintensity on in 11/37 (29.73%), lateralized to the left in 9/37 (24.33%), and T1W sequences corresponded to lesions in nerve roots 3/5, non-lateralized in 4/37 (10.8%) cases. intramedullary 1/5, and epidural 1/5, and hyperintensity was All the lesions were ill-defined. In relation to the spinal observed in 4/5 epidural lesions and 1/5 paravertebral mass. cord, lesions were intramedullary in 4/37 (10.81%) cases, In the last group, the paravertebral mass corresponded to the extramedullary in 32/37 (86.48%) cases, and both intra- and follow-up of a case formerly epidural which maintained the extramedullary in 1/37 (2.7%) case. Spinal cord compression same relaxometry. Frontiers in Veterinary Science 06 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 FIGURE 1 MRI studies of a 1-year-old domestic short-haired paraplegic cat. A multifocal large cell lymphoma with an intramedullary lumbosacral lesion and a sacrococcygeal epidural lesion was diagnosed. On T2W sagittal plane (A) the spinal cord is swollen with an increased heterogeneous signal from L2 to L6 vertebrae (arrows). On T1W post-contrast SPIR sagittal plane (B) there is a diffuse intramedullary contrast enhancement (white arrow) and an intense epidural enhancement from L7 to the first coccygeal vertebrae (gray arrow). On transverse planes at the level of L4-L5 vertebrae, the spinal cord appears swollen due to an intramedullary lesion with a right component which is hyperintense on T2W (C), isointense on T1W (D), and shows heterogeneous enhancement on post-contrast SPIR (E) images. On transverse planes at the level of the L7 vertebra, there is an epidural infiltrative lesion hyperintense on T2W and isointense on T1W images (gray arrows in F,G, respectively) with intense post-contrast enhancement (gray arrow in H). An increased signal of the L7 bone marrow with cortical sparing is observed (arrowheads in A,B,F,H). After chemotherapy and clinical improvement, an MRI was performed at 3 months (I–P). On the sagittal planes (I, T2W; J, T1W post-contrast SPIR) the L7 bone marrow signal was normalized (arrowheads) and a malacic lesion, presumptively residual, was present along the L5 vertebra (arrows). On transverse planes at the level of L4-L5 vertebrae (K, T2W; L, T1W and M, T1W post-contrast SPIR), the previous intramedullary lesion was not observed, and signs of malacia were seen (arrow in K). The corresponding L7 vertebra transverse planes (N–P) did not show epidural changes. STIR sequences were obtained in 14/27 cases. All of them Regarding the pattern of contrast enhancement it was showed a hyperintense and heterogeneous signal of the lesions homogeneous in 21/36 (58.33%) lesions and heterogeneous (Figure 3A) and corresponded with hyperintensity on T2W in 15/36 (41.66%) lesions. In the only case that an epidural images in all cases. Additional diffuse hyperintensity of bone was lesion and a separate intramedullary lesion, the first enhanced observed in 2 cases on presentation (affecting the L1 vertebral homogeneously whilst enhancement was heterogeneous in the body in one, and L5 and L6 vertebral bodies in other) and in 1 intramedullary lesion (Figures 1B,E,H). case in the follow-up (involving the right iliac crest, and the L6 Additional soft tissue involvement, apart from the paraspinal and L7 vertebral bodies case 2 in Table 3). masses, was found in 11/27 (40.7%) cases. These included 12 On T1W sequences after contrast administration, fat lesions, encompassing the paraspinal soft tissues contiguous to saturation (SPIR) was applied to 11/27 studies. Enhancement the lesions in 8/12 (66.6%) and distant soft tissue abnormalities was not observed in 1/37 (2.7%) lesions, for the rest in 4/12 (33.3%) with renal and intestinal enhancement (1/12, enhancement was intense (17/37, 45.94%), mild (14/37, 8.3%), hepatomegaly (1/12, 8.3%) and iliac adenomegaly (2/12, 37.84%), or moderate (5/37, 13.51%). In the only case with 2 16.6%), respectively. When present, the relaxometry of the intramedullary lesions enhancement was mild in one of the paraspinal soft tissue lesions was similar to the vertebral lesions and intense in the other, in the remaining cases if canal lesions. lesions were multifocal enhancement was similarly observed Local bone involvement was present in 7/27 (25.93%) in all. The non-enhancing lesion corresponded to a case cases, affecting multiple vertebrae and the sacrum in 2/27 with an epidural focal lesion that had previously received (7.41%) cases, and only one vertebra in 5/27 (18.52%) cases. In corticosteroids. In this case (case 3 in Table 3) when a addition, in one case, involvement of the L6 and L7 vertebrae second MRI was performed 3 months after surgery and and the right iliac wing was only present in the follow- chemotherapy the 2 new epidural lesions observed showed mild up MRI (case 2 in Table 3). When the bone was involved contrast enhancement. the MRI pattern was characterized by diffuse signal changes Frontiers in Veterinary Science 07 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 FIGURE 3 Dorsal STIR (A) and transverse T1W pre- (B) and T1W FIGURE 2 post-contrast (C) images of the lumbar vertebral column of an Cervical neurolymphomatosis in a 2.5-year-old domestic 11-year-old Turkish Van cat with progressive non-ambulatory short-haired cat which presented with a 3-month-course of paraparesis. There is bilateral diffuse nerve root enlargement right thoracic limb lameness progressing to monoparesis. In the (arrows), which shows hyperintense on STIR (A), mildly MRI transverse planes at the level of C6-C7 intervertebral hyperintense on T1W (B) and heterogeneously enhancing (C). A foramina, there is a diffuse enlargement of the right C7 spinal surgical biopsy smear (D) showed an abnormal population of nerve and root with mild spinal cord contralateral displacement pleomorphic lymphoid cells, with immature large lymphocytes (arrows on A–C). The lesion is iso-hyperintense on T2W images with more abundant cytoplasm and pale chromatin (arrows), (A), isointense on T1W (B), and enhances with an intense and Romanowsky, 100x. Intraoperative euthanasia was elected, on heterogeneous pattern on post-contrast SPIR images (C). On the macroscopic sample (E) the nerve roots from L5 to sacral biopsy tissue (D) neoplastic lymphoid cells are infiltrating spinal cord segments were enlarged and swollen (arrows). adjacent tissues, HE. CD 20-immune positive cells predominate in the neoplastic lymphoid cell population (E). The final diagnosis was neural and ganglionar small non-cleaved B-cell lymphoma. TP was determined in 10 cases and was elevated in 8/10. Specific measurements ranged from 30 to 300 mg/dl (mean value 103.4 mg/dl) in the samples retrieved from cervical and contrast enhancement in the bone marrow with cortical cisternal puncture and determination was 100 mg/dl in the sparing (Figures 1A,B,F–H, 5A–C) in all cases except one case sample obtained from a lumbar puncture. Albuminocytological (corresponding to an epidural focal lesion) in which a small dissociation was present in 3 cases. focus of cortical signal loss was observed with overall vertebral Corticosteroids were used previously to CSF sampling in shape preservation. 6/12 cats, in 1/6 results were normal, in 2/5 NCC was elevated in the lower limit (6 and 9 cells, respectively) and 3/5 corresponded to albuminocytological dissociation. CSF analysis CSF data were collected from 12/27 cases, results are summarized in Table 4. In one case the sample was obtained Cytopathological and histopathological by lumbar puncture and for the rest by cerebellomedullary studies cistern puncture. Cell count and/or cytology were abnormal in 8/12 cases. Pathological studies were performed in samples obtained NCC count was elevated in 7/12 cats (range 6–547 cells/microl, from the lesions by FNA (9/27), tissue biopsy (12/27), necropsy mean 224.5 cells/microl). Differential cytology was abnormal in (6/27), or a combination of them. Details about cell morphology, 6/12 cats. Cell count and cytology were normal in 4/12 cases. mitotic count, presence of necrosis, and IHQ are shown in A cell population different from 100% lymphocytes was Table 4. found in 4 cases. Neutrophils were present in 2 cases (with In the 9 cases in which FNA was performed the sample was 37 and 96%, respectively), and activated monocytes were taken from epidural masses in 6/9 cases and from perivertebral present in 2 cases (7 and 2%, respectively). Morphologically affected tissue in 3/9 cases. Additional FNA of a cutaneous abnormal lymphocytes were observed in 5 cases, with activated nodule was performed in 1 case. In 9/9 results were compatible lymphocytes present in one case and a population of large with lymphoma with a high population of morphologically neoplastic lymphocytes in 4 cases, one of them further atypical lymphoid cells. In the case sampled from an epidural categorized as large granular cell lymphoma. lesion and a cutaneous nodule, results were similar from both Frontiers in Veterinary Science 08 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 samples. A further description was available in one case with a Treatment and outcome paravertebral mass that was characterized as a presumptive large granular cell lymphoma and later confirmed by necropsy. Results from treatment and survival time are summarized A surgical tissue biopsy was performed in 12/27 cases. in Table 6. In 8/27 cases surgery for decompression and tissue From vertebral canal masses in 11/12 [epidural localization sampling for biopsy was performed and was followed by 5/12 (Figures 5, 6), paravertebral with vertebral canal invasion chemotherapy in 4/8. Chemotherapy alone was the treatment for 2/12, or nerve roots 4/12] including surrounding tissues (bone, 2/27 cases. Euthanasia was performed on the day of the MRI in epidural fat), and a paravertebral lesion in 1/12. An additional 4/27 cases, and in <1 week in 6/27 due to worsening of the signs biopsy of a cutaneous nodule was obtained in 1 case. For the despite medical treatment, 1/27 died in the following days. Data cases with epidural lesions infiltration to bone was reported in 3 from survival time since MRI longer than 1 week were available of them. for 9/27 cats and ranged from 1 week to 13 months (mean 4.5 Cytology from an intrasurgical lesion smear was months), but 5/9 cats were lost to follow-up at that recorded performed in 3/12 cases (Figure 3D) with results time, therefore the last known survival time was considered for compatible with lymphoma in all, in one case further the study. For the rest of the 7/27 cats survival time is unknown categorization as intermediate-large cell was achieved by as they were lost to follow-up after diagnosis. histopathological studies. For the 2/27 cases treated with chemotherapy alone, A lymphocytic proliferative population of cells was present survival time was 4 months (epidural intermediate/large cell in all samples. IHQ was performed in 8/27 cases, in 5/9 there lymphoma) and 12 months (epidural and intramedullary large was immunopositivity for T-cells and in 3/9 for B-cells. cell lymphoma), respectively. For the 4 cases managed with Ten cats were euthanized either after MRI diagnosis surgery without chemotherapy survival time was recorded for or in the following days due to rapid clinical worsening, 2/4 and was 2 weeks (epidural large B-cell) and 13 months corresponding to epidural focal (3/10), epidural multifocal (epidural small non-cleaved T-cell, at that time a new lesion (1/10), epidural/intradural/intramedullary (1/10), nerve with paravertebral location was diagnosed, case 2 on Table 3), roots (2/10), paravertebral mass (2/10) and epidural respectively. For the 4 cases treated with surgery combined and paravertebral mass (1/10) localizations. One cat with with chemotherapy time of survival was at least 1 month for multifocal intramedullary lesions died in the following days 2/4 cats (epidural intermediate/large B-cell and small B-cell after diagnosis. nerve roots neurolymphomatosis), 3 months for 1/4 (epidural Anatomopathological studies were performed in 6/27 cases. intermediate/large cell, at that time MRI showed new epidural On 2/6 multicentric involvement of other organs (liver, spleen, lesions, case 3 on Table 3) and at least 6 months for 1/4 lungs, and/or intestines) was observed. The studies included (paravertebral large cell). 3/6 cases with epidural localization based on MRI, 1/6 Data regarding survival time was available for 14 of the cases intramedullary, 1/6 paravertebral mass with vertebral canal in which tissue was obtained either by biopsy or necropsy. In this invasion, and 1/6 multifocal case with epidural and paravertebral group for the 5/14 cases that showed necrosis in the pathological masses with vertebral canal invasion (Figure 4). studies, 4/5 (80%) died or were euthanized ≤ 1 month after the In general, an infiltrative pattern was described in all the MRI. For the rest of the 9/14 cases that did not show necrosis tissue samples, with the presence of necrosis and/or hemorrhage 4/9 (44.44%) died or were euthanized ≤ 1 month after the MRI. reported in 5 cases. Cell pleomorphism was reported in 10 Results are shown in Table 7. cases. Mitotic count was recorded in 16 cases and described as low (7/16), moderate (3/16), or high (6/16). According to histopathology tumors were classified as low grade (1/27), Statistical analysis intermediate grade (22/27), and high grade (4/27). Data are summarized in Tables 4, 5. When data from all the MRI studies were retrieved explanatory variables were obtained regarding lesion localization (cervical, thoracic, lumbar/lumbosacral), Diagnosis lateralization (circumferential, right, left, no), extension (≤ or >1 vertebral body), compression degree (mild, moderate, Final diagnoses in this group of 27 cats were achieved by CSF severe), T2W signal (iso, iso-hyper, hyper), T1W signal (iso, (4/27, 14.8%), FNA (6/27, 22.2%), FNA and histopathology on iso-hyper, hyper), contrast enhancement (none, mild, moderate, surgical biopsy (2/27, 7.4%), cytology from smears on surgical intense), contrast enhancement pattern (homogeneous, biopsy (2/27, 7.4%), cytology from smear and histopathology on heterogeneous), vertebral-bone involvement (present/absent) surgical biopsy (1/27, 3.7%), histopathology on surgical biopsy and extraneural findings (present/absent). In the chi-square (6/27, 22.2%), necropsy (5/27, 18.5%), or CSF, FNA, surgical goodness of fit test significant (p-value < 0.05) values were biopsy and necropsy (1/27, 3.7%). obtained for a higher prevalence of lumbar/lumbosacral Frontiers in Veterinary Science 09 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 FIGURE 4 MRI of an 8-year-old domestic short-haired paraplegic cat diagnosed with multicentric small-intermediate T-cell lymphoma. Sagittal plane T1W post-contrast SPIR of the caudal thoracic and lumbar vertebral column (A) shows a paravertebral mass ventral to the L1 vertebra and extending cranially to T12 and T13 vertebrae (arrows) and irregular hyperintensity on the bone marrow of several vertebral bodies (arrowheads). Transverse planes at the level of T13-L1 intervertebral space (B–D) show an extension of the paravertebral mass through the intervertebral foramina (arrows) and to the epidural space circumferential to the spinal cord. The lesion is hyperintense on T2W (B), isointense on T1W (C), and enhances with a mild heterogeneous pattern on post-contrast T1W (D) images. Transverse planes T2W (E) and post-contrast T1W (F) at the level of the L6 vertebra, show an epidural diffuse infiltrate (arrows) with similar signal features. Hyperintensity of the vertebral body with cortical sparing is also depicted (arrowhead in F). On histopathology image, HE (G) neoplastic lymphoid cells are surrounding a nerve root, on IHC (H) T-lymphocytes, CD3-immune positive cells, predominate in the tumor. FIGURE 5 Intermediate-large B-cell lymphoma in an 8-month-old domestic shorthair cat with a history of progressive paraparesis and caudal lumbar spinal pain for 10 days. On MRI transverse planes at the level of L6 (A–C), there is an epidural lesion isointense on T2W (arrows in A) and T1W (arrows in B) images. On T1W post-contrast SPIR (C) there is an intense enhancement of the lesion (arrows) as well as diffuse vertebral bone marrow enhancement. The spinal cord appears severely compressed with dorsal displacement (arrowheads in A–C). On HE stain of tissue biopsy (D), there are neoplastic lymphoid cells, mixed with erythrocytes, infiltrating nervous tissue. B lymphocytes, CD20-immune positive cells, predominate in the tumor in IHC (E). (p-value = 0.01171) and circumferential (p-value = 0.04906) presence of necrosis on histopathology as a dependent localization, T2W hyperintensity (p-value = 4.364e-06), T1W variable, the only significant results were found for isointensity (p-value = 8.998e-07), and bone involvement T2W imaging. A T2W hyperintensity was significantly (p-value = 0.001177). correlated with the non-presence of necrosis (coefficient When the explanatory variables data from the MRI −2.084253, p-value = 0.017) when compared to T2W features were converted to binary and compared to the iso-hyperintensity. For T2W isointensity the estimator Frontiers in Veterinary Science 10 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 FIGURE 6 MRI of a 4-year-old domestic short-haired cat with progressive ambulatory paraparesis for 7 days, diagnosed with large B-cell lymphoma. Transverse MRI of a cat at the level of the cranial portion of the L3 vertebral body (A–F). There is an epidural dorsolateral right mass (arrows on A–C) isointense on T2W (A) and T1W (B) images, which shows intense enhancement on post-contrast T1W SPIR image (C) and induces mild spinal cord contralateral displacement. The images (D–F) correspond to the same plane and sequences as (A–C), respectively, 18 days later after a right hemilaminectomy for decompression and tissue sampling was performed. A left epidural lesion was found (arrows on D–F) inducing a greater spinal cord compression. Tissue biopsy (G,H) HE, neoplastic lymphoid cells including basophilic (dark blue) necrotic foci and mixed with infiltrating erythrocytes. was not significant for a 5% confidence level therefore In the present study distant lesions on abdominal organs there do not seem to be differences when compared were depicted in 4/27 of the cases, compatible with multicentric to iso-hyperintensity. lymphoma. Alterations were present on MRI in 2 cases, on necropsy in 1 case, and in the other case, MRI findings of multicentric pattern were confirmed by necropsy (Figure 4). Due to the low number of complete necropsies (5/27), the actual Discussion incidence of multicentric lymphoma cannot be inferred but may Primary and secondary incidence be greater, especially taking into account the previous studies. It is also possible, as it has been pointed out in another study It has been reported that 90% of feline lymphomas affecting (6) that some cases could be secondary although no other sites the CNS have a spinal cord localization (4), with an overall of lymphoma were detected in the diagnostic work-up, which prevalence of primary lymphoma considered to be low (19), would explain the predominantly extramedullary or combined similar to the human counterpart (11). Spinal cord lymphoma intra- and extramedullary distribution of the lesions in our is most commonly secondary, encountered as a part of a group. On the other hand, it should be considered that feline multicentric process, frequently with renal or bone marrow lymphomas, when localized in the spinal cord, can be primary. involvement (20). In one study multicentric involvement was present in 84.6% of cats with spinal cord lymphoma (2), and another study showed that although all cats were evaluated Age and sex for neurologic deficits, 85% of the necropsied cases had a multicentric disease (21). Primary lymphoma, originated and A bimodal tendency was observed in the present study with restricted to the spinal cord, has been also described in cats with 40.7% of cats ≤2.5 years and 44.4% of them aged between 8 and intramedullary and epidural localizations reported (6, 19, 22). 15 years. Spinal cord lymphoma was traditionally considered to In humans, almost 100% of primary spinal cord lymphomas be a juvenile neoplasm in cats (4, 21) but has been reported as affect intramedullary structures and the majority of metastatic well as one of the most common spinal cord diseases in cats > lymphomas are extramedullary (6). It is worth remarking that 2 years of age, with median and mean ages of 4.5 and 6.3 years, in previous literature epidural lymphomas have been often respectively (1). A bimodal age has been reported before in feline grouped with CNS lymphomas, however, considering that the lymphoma, with a group of cats younger than 3 years and a epidural space does not belong to the CNS, the term CNS group of cats older than 8 years (23). A recent report showed lymphoma applied for general cases or cases affecting the similar results with a wide age distribution ranging from 1 to epidural space should be avoided. 20 years, with a median age of 14 years (7). In light of the strong Frontiers in Veterinary Science 11 frontiersin.org Frontiers in Veterinary Science Lorenzo et al. TABLE 4 Results of CSF analysis and cytopathological studies in 27 cats with spinal lymphoma. CSF FNA Biopsy Necropsy Cell Mitotic Necrosis IHQ Pathological (12/27) (9/27) (12/27) (6/27) morphology count diagnosis Epidural focal - 1 - - Large - - Large cell (12/27) - 1 - - Large - - Large cell - 1 - - Large - - Large cell - 1 - - Large - - Large cell - - 1 - Large, pleo morphic Moderate Yes B-cell Large B-cell - - - 1 Interm, pleo morphic Low - - Intermediate cell - 1 - - Interm/large, - - - Interm/large cell pleomorph ACD - 1 - Interm/large, Moderate - - Interm/large lymp monomorp Normal - 1 - Interm/large, Low - B-cell Interm/large B-cell pleomorph - - 1 - Small High Yes T-cell Small non-cleaved 12 T-cell Normal - - 1 Interm-large, Low - - Lymphoblastic monomorp - - - 1 Small, mon omorphic High - - Small non-cleaved cell Total 3 5 4 3 Epidural multifocal – 1 1 - Large Low - - Epitheliotropic T-cell (1/27) Epidural and Neoplastic cells - - - Large - - - Large cell intramedullary multifocal (1/27) Epidural/intradural/ Neoplastic cells - - - Large, pleomorphic Large granullar cell intramedullary (1/27) 10.3389/fvets.2022.980414 Intramedullary ↑NCC, ↑TP - - 1 Pleomorph, High Yes Leptomeningeal multifocal (1/27) occasional lymphoplasmacytic plasmacytoid lymphomatosis differentation frontiersin.org (Continued) Frontiers in Veterinary Science Lorenzo et al. TABLE 4 (Continued) CSF FNA Biopsy Necropsy Cell Mitotic Necrosis IHQ Pathological (12/27) (9/27) (12/27) (6/27) morphology count diagnosis Spinal nerve Neoplastic - - - Large - - - Large cell roots/intracanal (5/27) Neoplastic - - - Large - - - Large cell - - 1 - Large - - - Large cell - - 1 - Interm-large, - - - Interm/large pleomorph ↑NCC - 1 - Small Low - B-cell Small non-cleaved B cell Intramedullary and ↑NCC, ↑TP - 1 – Medium-large, High - T-cell Interm/large nerve roots multifocal pleomorph (1/27) 13 Paravertebral mass ↑NCC, ↑TP 1 1 1 Large High Yes T-Cell Large granular T-cell, and epidural focal multicentric (4/27) - 1 1 - Large, monomorphic High - - Large cell - 1 - - Large, pleomorphic Low - - Large cell ACD - 1 - Small-interm, Low - T-cell Small/interm T-cell, pleomorphic multicentric Epidural and - - - 1 Small/interm, Moderate Yes T-cell Multicentric paravertebral mass pleomorphic lymphoma multifocal (1/27) Total Normal 2/12 Small: 2 Abnormal 6/12 Small-medium: 2 Diagnostic 4/12 Medium: 1 10.3389/fvets.2022.980414 Medium-large: 6 Large: 14 frontiersin.org CSF, cerebrospinal fluid; NCC, nucleated cell count; TP, total protein; ACD, albuminocytologycal dissociation; FNA, Fine needle aspirate. Lorenzo et al. 10.3389/fvets.2022.980414 TABLE 5 Frequency of distribution in 27 cases of feline spinal Clinical signs lymphomas classified according to the National Cancer Institute working formulation (NCI WF). In this group of cats, the presentation was generally acute to Grade Tumor type Number (%) sub-acute with rapid progression. For most of the cats (90.4%) clinical signs were present from 1 day to 1 month with the Low Lymphocytic plasmacytoid 1 majority of them (57.1%) having a duration of ≤7 days. For the Total 1/27 (3.7) rest, the duration of signs did not exceed 3 months. Intermediate Mixed cell 8 The most common reported initial clinical sign, affecting Large cell 14 96.3% of all the cats, was progressive paresis (85.2% paraparesis, Total 22/27 (81.5) 7.4% tetraparesis, 3.7% monoparesis). Besides, lameness was the High Small non-cleaved cell 3 initial complaint in 3.7% of cases. Lymphoblastic 1 The clinical signs observed when cats presented at the Total 4/27 (14.8) referral centers included signs of spinal pain (33.3%) and neurological deficits referred to a spinal cord localization most commonly affecting the pelvic limbs (85.1%). Other signs included tetraparesis (7.4%), thoracic limb monoparesis (3.7%), bimodal distribution, neither the mean nor median value depicts and tetraplegia (3.7%). a real age predisposition. Similar presenting signs have been previously reported, with No sex predisposition was found in this study, as male and a duration of <60 days (2). The rapid progressive neurological female cats were similarly affected. A study of incidence and risk deterioration, in some cases after an initial insidious course, was factors for all types of feline lymphoma found that male cats had in concordance with other reports (4). an increased risk compared to females similar to other studies, but others, as well as the present study, have failed to report an association between lymphoma and sex (24). MRI distribution and pattern In our group of 27 cats with spinal lymphoma, an epidural solitary lesion was the most common localization at presentation Relation to FeLV (44.4%). Other localization patterns varied and mostly included vertebral canal nerve roots (18.5%) or a paravertebral mass In this study, FeLV test results were negative for 57.1% and with vertebral canal invasion (14.8%), with other localizations positive for 42.9%. This prevalence is in between the previously less commonly seen. When the 27 cases plus the 3 follow-ups referenced values, which shows variations depending on the that had lesions were considered together, lesions were focal studies. Cats have been reported to be positive for FeLV in 17.6% in the majority of the cases (80%), for the rest, a multifocal of cases in a study of feline intracranial lymphoma in 2003 (25) localization with more than one lesion affecting the nervous and in 56.5% of cases in a study of spinal cord lymphoma in 2008 system was observed. Multifocal lesions affecting different spinal (2). Since 1990 a decrease in the prevalence of FeLV infection cord compartments were found in 6.66% with localizations in cats with lymphoma has been described, probably related epidural and intramedullary and intramedullary and nerve to vaccination and population control programs (20, 26), and roots. All the vertebral canal masses had indistinct (ill-defined) nowadays the majority of cats with lymphoma are unaffected by margins. Extensive vertebral canal lesions spreading more than FeLV so the clinical importance of this factor is comparatively 1 vertebral body were present in 64.86% but this variable was not low (9). In the study by Meichner et al. in (27), only 13% of cats statistically significant (p > 0.05). with lymphoma were FeLV antigen-positive. However, Mello Similarly, previous reports have described the MRI et al. reported in 2019 a rate of cats with NS lymphoma infected characteristics of spinal cord lymphoma in cats showing similar with FeLV of 87.5% similar to proportions presented prior to the features to those seen in dogs (7) and to the intracranial 1990s, which may reflect differences in prevalence in different counterpart (6), as lesions commonly indistinctively marginated regions of the world. from the surrounding tissue, mostly with a focal (80.6%) and The bimodal age has been related to FeLV in feline extensive (64.5%) pattern (6, 8). Others have also reported lymphoma, with cats younger than 3 years widely associated spinal lymphoma as focal, from 76% (21) to 95.6%, presented with FeLV infection and cats older than 8 years generally free as a single lesion that may expand over more than one vertebral from infection (23). The same tendency seems to be present in body, or multifocal, as two or more separate lesions (4). In a this study, as for the group of younger cats most of them (63.6%) report of 8 cats with spinal cord lymphoma (23) the pattern were FeLV positive and for the group of older cats the majority was also often focal (75% of cases). In a similar way, it has (83.3%) were negative. been reported a tendency of spinal cord lymphoma to form Frontiers in Veterinary Science 14 frontiersin.org Frontiers in Veterinary Science Lorenzo et al. TABLE 6 Treatment and survival time in 27 cats with spinal lymphoma. Number Surgery Chemo- Medical Euthanasia Euthanasia Survival time Final pathological Grading (NCI (%) therapy management on the day on the (known) diagnosis WF) of MRI following days of MRI Epidural focal 12/27 - - 1/2 - ? ? Large cell Intermediate (44.4) - - 1/2 - ? ? Large cell Intermediate - - 1/2 - ? ? Large cell Intermediate - - 1/2 - - 1 week, LTF Large cell Intermediate 1/12 - 1/2 - - 2 weeks Large B-cell Intermediate - - - 1/2 - - Intermediate cell Intermediate - 1/12 - - - 4 months, LTF Interm/large cell, pleom Intermediate 1/12 1/12 - - - 3 months Interm/large, monom Intermediate 1/12 1/12 - - - 1 month, LTF Interm/large B-cell Intermediate 1/12 - 1/12 - - 13 months Small non-cleaved T-cell High - - 1/12 - 1/12 - Lymphoblastic High - - 1/12 - 1/12 - Small non-cleaved cell High 15 Total 4/12 3/12 8/12 1/12 2/12 Epidural multifocal (3 1/27 (3.7) - - 1/1 - 1/1 - Epitheliotropic T-cell Intermediate lesions) (metastatic?) Epidural and 1/27 (3.7) - 1/1 - - - 12 months Large cell Intermediate intramedullary multifocal (2 lesions) Epidural/intradural/intra- 1/27 (3.7) - - - 1/1 - - Large granullar cell Intermediate medullary Intramedullary 1/27 (3.7) - - 1/1 - 1/1 (death) - Leptomeningeal Low multifocal (2 lesions) lymphoplasmacytic lymphomatosis Nerve roots 5/27 (18.5) - - - 1/5 - - Large cell Intermediate intracanal 10.3389/fvets.2022.980414 - - 1/5 - - ? Large cell Intermediate - - 1/5 - - ? Large cell Intermediate - - 1/5 - 1/5 - Interm/large cell Intermediate frontiersin.org 1/5 1/5 - - - 1 month, LTF Small non-cleaved B cell High (Continued) Lorenzo et al. 10.3389/fvets.2022.980414 TABLE 7 Survival time at 1 month in 14 cats with or without necrosis observed in tumoral tissue samples. Grading (NCI Survival Survival Number Intermediate Intermediate Intermediate Intermediate Intermediate Intermediate time ≤ 1 time > 1 (%) WF) month month Necrosis 4/5 (80) 1/5 (20) 5/14 (35.7) Final pathological No necrosis 4/9 (44.44) 5/9 (55.55) 9/14 (64.3) Large granular T-cell Small/interm T-cell Multicentric T-cell Number (%) 8/14 (57.14) 6/14 (42.86) Large monom Interm/large diagnosis Large pleom extensive lesions (21) with most tumors extending over multiple vertebral bodies, and 19% with more than 1 level of spinal cord Survival time 6 months, LTF involvement. Other studies had reported that in one-half of the (known) cases spinal lymphoma extends to other spinal cord segments (2, 5). ? ? The selection of cases of the present study implied vertebral - - - canal invasion, and in relation to the spinal cord lesions days of MRI Euthanasia following were most frequently extramedullary (86.48%), followed by on the 7/27 1/4 1/4 intramedullary (10.81%), and less frequently both intra- and - - - - extramedullary (2.7%). The degree of spinal cord compression of extramedullary lesions most commonly ranged from moderate (39.39%) to severe (36.36%). Previously, lymphoma of the Euthanasia on the day spinal cord has been reported as extradural, intradural– of MRI 4/27 1/1 extramedullary, intramedullary, or with a combination of these - - - - - compartments (20). A previous study (2) revealed extradural and intradural management components in 29 of 33 (87.9%) cats with lymphosarcoma, Medical with 34.5% having an exclusive extradural location and 10.3% 16/27 1/4 1/4 1/4 - - - an exclusive intradural location. When extradural, lymphoma has been described typically as an epidural mass resulting in spinal cord compression, which can extend into and through the adjacent vertebrae as well as into the underlying muscle Chemo- therapy 6/12 (7). The lesions may have a variable mass effect from mild to 1/4 - - - - - NCI WF, National Cancer Institute working formulation; LTF, lost to follow up. moderate (2, 6, 7). Similarly to our study, it has been reported predominantly situated extradural to the spinal cord (20, 23), moreover, a study found that when the lesion was centered on Surgery epidural soft tissues, it was more likely lymphoma (7) and rarely 8/27 1/1 1/4 1/4 - - - occurs as an intramedullary tumor (2, 8). On the other hand, in a study of 92 cats with clinical signs of spinal cord disease (3) 3/7 of cats diagnosed with spinal lymphoma had intramedullary lesions. Intradural-extramedullary lymphoma occurs as a non- Number 4/27 (14.8) 1/27 (3.7) 1/27 (3.7) delineated mass replacing the subdural and subarachnoid spaces, (%) and commonly infiltrating the spinal cord parenchyma or nerve roots (1, 20), but the pattern of intradural infiltration without an TABLE 6 (Continued) epidural mass was not found in our study. In a recent study (8) nerve roots (2 lesions) multifocal (2 lesions) lesions frequently affected multiple spinal compartments similar Intramedullary and Paravertebral mass paravertebral mass and epidural focal to previous publications (1, 2) conversely, although intradural- Epidural and extramedullary involvement with epidural lesions could be observed on MRI, both intra- and extramedullary involvement Total were present only in one lesion on this study. In humans, an Frontiers in Veterinary Science 16 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 epidural location for lymphoma is observed in 0.1–6.5% of all found. Therefore, it may be concluded that recurrence of lymphomas (11). spinal lymphoma with the appearance of different localizations Regarding localization in relation to the vertebral column, is possible. lesions were significantly most frequent in the lumbar and Regarding the patterns of signal intensity, on T2W lumbosacral areas (51.35%, p-value < 0.05). Other localizations sequences the lesions showed mainly hyperintense (67.57%, p- included the thoracic (35.13%) and cervical (13.51%) segments. value < 0.05) isointense (18.91%), or iso-hyperintense (13.51%). In previous studies, spinal cord lymphoma exhibited a On T1W sequences the most prevalent signal was isointense predilection for lumbar (1, 8, 23) and thoracic regions (4, 8) (72.97%, p-value < 0.05), the remaining lesions showed iso- or both (2, 7, 20). It has been observed that as lumbosacral hyperintense (13.51%) or hyperintense (13.51%). On STIR spinal cord segments are generally affected by lymphoma, this is sequences lesions were hyperintense, corresponding to the T2W frequently associated with more spinal cord region involvement sequences. In a follow-up study (Figure 6) the new lesion (1, 20) but this tendency was not found in this report. In had an increased signal on T2W images compared to the humans, the most common region of involvement is the thoracic previous lesion. vertebral column, followed by the lumbar and cervical vertebral In previous studies, a general preponderance of column (11). hyperintensity on T2W sequences and isointensity on Nerve root involvement was present in 22.2% of cases, with T1W sequences has been reported as well. In a study (7) cervical (50%), lumbar (33.33%), and lumbosacral (16.66%) the percentage of T2W hyperintense extradural lesions was localizations. In one of the cervical cases, an intramedullary even higher and on T1W images lesions were predominantly lesion was also present, thus the case was considered multifocal. isointense (62.5%) or either hyperintense in a higher percentage It has been previously described that lymphoma may involve (37.5%), but this study included feline and canine spinal the spinal cord from vertebral canal lesions centered on the lymphomas. Similarly, a study with 2 cases (6) described nerve roots or from a vertebral canal extension along peripheral the lesions as T2-hyperintense and T1-hypointense in both nerves (8, 20, 28), but lymphoma of the peripheral nerves, intramedullary and epidural lesions, and in another study (8) named neurolymphomatosis, has been considered sporadic the lesions were also described as hyperintense in T2W and in cats, and less frequently with vertebral canal invasion isointense on T1W. The other less common mixed patterns of and spinal cord involvement either by compression or/and relaxometry found in this study, iso-hyper on T2 or T1 have not infiltration of the subarachnoid space and spinal (4, 8, 21, been reported before. 26). Neurolymphomatosis of the lumbosacral plexus, although Enhancement of the lesions after contrast administration similar to our study less commonly reported than in the brachial was variable, and appeared as intense (45.94%), moderate plexus, was the most common lesion pattern in a group of (13.51%), or mild (37.84%). In the only case with 2 cats (8). MRI findings (8, 29) of lumbosacral lymphoma have intramedullary lesions enhancement was mild in one of the been described with symmetrical or asymmetrical thickening lesions and intense in the other, but in the remaining cases and enhancement of the nerve roots or/and nerves, meningeal if lesions were multifocal enhancement was similarly observed enhancement of the conus medullaris, with or without the in all. Contrast enhancement was not observed only in one presence of an extradural, intradural, or intramedullary mass, case with an epidural focal lesion. Regarding the pattern of or some combination of these. Conversely, in the study from contrast enhancement, it was either homogeneous (58.33%) Durand et al. that included 8 cats with spinal cord and spinal or heterogeneous (41.66%). Previous reports showed that nerves lesions, >50% of the cases showed lesions involving both, in general, contrast enhancement of the lesion is present which was more common than the present and previous reports. and has been described as homogeneous (6, 7, 20), but it A paravertebral mass with vertebral canal invasion was can be faint to none (8). The absence of enhancement has observed in 4/27 (14.8%) cases, with 3/4 located in the thoracic been also described in humans, as primary CNS lymphoma segments and 1/4 thoracolumbar. In the latter, a multifocal can present as longitudinally extensive transverse myelopathy presentation with a separate epidural lumbosacral lesion was and may be non-contrast enhancing on gadolinium-enhanced observed (Figure 5). Spinal lymphoma with paravertebral soft MRI (30). Our study shows similar results, as lesions most tissue involvement in the form of paraspinal masses has been commonly enhanced homogeneously, but not significatively previously reported (20) but not frequently described on MRI. In different from a heterogenous pattern. A previous study the study of Durand et al., the only case described corresponded reported that in cases with multifocal lesions, these lesions to a cat with multifocal hypaxial mass and epidural lesions. In the showed overall similar MRI characteristics (8), this observation study of Auger et al., in a group of 20 dogs and 5 cats, round cell is similar to our study, although one case showed 2 separate tumors were centered on paraspinal soft tissues in 8% of cases, epidural and intramedullary lesions with homogeneous and which is less than in our study. heterogeneous enhancement, respectively. A characteristic Interestingly, in the cases in which MRI was repeated contrast enhancement for spinal lymphomas was not found in new lesions with different locations and extensions were this study due to the variable degree and pattern. Frontiers in Veterinary Science 17 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 It has been reported that corticosteroids may reduce in lymphoma, with absent or mild cortical lysis with preserved or eliminate abnormal contrast enhancement which may vertebral shape. This pattern is similar in dogs (7) and humans complicate the interpretation of enhanced MRI (12, 13). Marked (31). In the study from Auger et al. bone lesions in cats and reduction in enhancement or clinical improvement following dogs showed different patterns of signal intensity on T1W corticosteroid therapy, although seen in other conditions, has and T2W and diffuse STIR hyperintensity. Failure of bone been suggested to be highly suspicious for CNS lymphoma marrow suppression on STIR sequences has been reported in (13). In a study of radiographic findings of primary CNS dogs with spinal lymphoma (32) and has been attributed to the lymphoma in humans nearly all the lesions enhanced, except replacement of normal bone marrow fat by malignant neoplastic after corticosteroid administration (10). In our study, 12 cats cells, inflammation, and/or necrotic tissue. In agreement with had received corticosteroids before imaging and all except one previous studies performed in dogs, a STIR sequence, either of this group showed contrast enhancement, therefore although in the sagittal or dorsal plane, should be part of the imaging corticosteroids may have influenced the degree of intensity, protocol of the vertebral column, especially in low-field magnets contrast enhancement of the lesions was generally observed. and if round cell neoplasia is a consideration (32, 33). On high In fact, the non-enhancing lesion corresponded to a case that field magnets, T1W contrast-enhanced fat-suppressed images had previously received corticosteroids. In this case (case 3 in demonstrate similar patterns to STIR (34). Table 3) when a second MRI was performed 3 months after In conclusion, as well as in humans (11, 30) a presumptive surgery and chemotherapy the 2 new epidural lesions observed antemortem diagnosis of spinal lymphoma can be established showed mild homogeneous contrast enhancement. Our study is based on MRI data. the first to report follow-up MRI studies, and therefore the first to report a change in the degree of enhancement. Extraneural soft tissue involvement, apart from the CSF and tissue studies paraspinal masses, was found in 40.7% of cases with a total of 12 lesions, affecting mainly the paraspinal soft tissues In our group of cats, the antemortem diagnosis was reached contiguous to the lesions (66.6%) but also with distant soft tissue in 81.5% of cases, which is above than reported before, and abnormalities (33.3%) with renal and intestinal enhancement was achieved either by CSF (14.8%), FNA (22.2%), surgical (8.3%), hepatomegaly (8.3%) and iliac adenomegaly (16.6%), biopsy (37%), or by a combination of them. In a study of feline respectively. When present, the relaxometry of the paraspinal lymphoma including intracranial, spinal, and spinal nerves soft tissue lesions was similar to the vertebral canal lesions. In lesions (8), antemortem diagnosis was achieved in 32.3% of one case with an epidural focal lesion, extraneural soft tissue cases, and similarly to our study was based on CSF cytology involvement appeared only in the follow-up MRI (case 2 in (4/10), histopathology on biopsy (3/10), tissue cytology (2/10), Table 3). Different from our study, extra-neural tissue invasion or a combination of cytology and histopathology (1/10). In was previously reported to be less common (16.7%) in cats with the same study lymph node FNA was performed in 4 cats, spinal lesions (8). On the other hand, abdominal lymph nodes being diagnostic for lymphoma only in 1 cat which had were the most frequently affected in secondary lymphomas diagnostic CSF cytology, therefore lymph node FNA did not in another study (23), and the presence of abdominal lesions add relevant information for diagnosis. Also, a lower percentage has been reported to be more likely in patients with round cell of antemortem diagnosis was reported in another study of neoplasia (7). 85 cases of tumors affecting the spinal cord of cats with an Local bone involvement was present in 26% of cases antemortem diagnosis of lymphoma reached in 21.2% of cases at presentation, affecting multiple vertebrae and the sacrum (2). In another study, the diagnosis of CNS lymphoma was based (7.41%) or only one vertebra (18.52%). In addition, in one on the identification of lymphoblasts in CSF in 3 dogs and 2 cats, case, involvement of the L6 and L7 vertebrae and the ilion and necropsy in 5 dogs and 2 cats (6). was only present in the follow-up MRI. Bone involvement was centered in the bone marrow and characterized by diffuse signal changes and contrast enhancement, better depicted on STIR CSF and post-contrast fat-saturated sequences. Cortical sparing was typically maintained, in only one case a small focus of cortical In our study, all the CSF parameters were normal only in signal loss was observed but anyway with overall vertebral shape two of the cases in which it was analyzed. Differential cytology preservation. In cats diagnosed with lymphoma in previous was abnormal in 50% of cats and in 33% of cases neoplastic studies (7, 20, 23), some cats had polyostotic bone lesions with large cells were observed therefore allowing a diagnosis. In one cortical sparing on MRI, with multifocal sites of extradural case small lymphocytes were observed, but it was not possible spinal cord compression, whereas others had lesions centered on to give a definitive diagnosis based on cytology morphology. the epidural soft tissues with no evidence of bone involvement. As the differentiation between small neoplastic and reactive A previous study (8) has shown a similar characteristic pattern lymphocytes may be difficult in feline lymphoma (14) additional Frontiers in Veterinary Science 18 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 test as flow cytometry (13) or Polymerase chain reaction for atypical lymphoid cells as monoclonal (neoplastic). As well as antigen receptor rearrangement (PARR) (8), may assist in the in animals, there is some overlap in the morphologic features of diagnosis but those were not performed in the present study. neoplastic and inflammatory lymphocytes, and this can make Total NCC count was elevated in 58% of cases with the interpretation of cytology difficult. The same study indicates neoplastic lymphocytes observed in 3 of them. It has been that the sensitivity of CSF cytology varies widely (2–32%). reported that CSF was diagnostic for lymphoma in the cases with It has to be considered that the previous use of higher NCC counts which is in agreement with the previously corticosteroids may have influenced the CSF analysis results, reported difficulty in diagnosis in CSF with low cells number especially regarding the NCC. Also, the sensitivity of CSF (13). In a study of inflammatory CSF in 62 cats (35), CSF analysis cytology is reduced after exposure to corticosteroids (13). alone was useful in the diagnosis of cats with lymphoma. In this In the present study, in the 6 cats previously treated with study, all cats had mononuclear inflammation, and in 5 cats large corticosteroids in which CSF was sampled the result was normal atypical lymphoid cells were observed in the CSF. Similarly to in 1/6, the number of cells was elevated but low in 3/6 (range our study, total NCC was variable (6–1,144 cells/µl). 6–9 cells/microL) and albuminocytological dissociation was TP was determined in 10 cases and was elevated in found in 2/6. In these cases, the NCC could have been lowered 80% with a wide range of 30–300 mg/dl. A wide variation by the corticosteroid effect. The small amounts of the samples has been reported as well in cats with lymphoma with TP generally obtained from cats may have also influenced the levels varying from normal to elevated (31–1,000 mg/dl) (35). ability for diagnosis. In humans, sensitivity improves when a Albuminocytological dissociation was present in 30% and larger volume (≥10.5 mL) is analyzed but this is not feasible corresponded to 2 cats with epidural lesions and 1 cat with a in cats. Again in humans, when CSF results are inconclusive paravertebral mass. repeated CSF examinations are indicated for definitive diagnosis CSF examination remains a good diagnostic option in cases (13, 30, 36). Serial CSF samples should be considered in of lymphoma with intradural invasion (20), with a diagnosis cases with compatible MRI findings if FNA of the lesion/s supported by the identification of large neoplastic lymphocytes is not possible, and also as a less invasive procedure than (6, 14). It is worthwhile to point out that in the veterinary surgical biopsy. literature, it has been common to refer to the largest malignant According to the present and previous (8) studies, the MRI lymphocytes as lymphoblasts, whereas in human pathology, the features when combined with CSF analysis may provide an term lymphoblast is used to describe an immature cell of small antemortem diagnosis in cats with spinal lymphoma. size, larger than a mature lymphocyte and smaller than a large lymphocyte, and this nomenclature should also be used (16). Close to our results, in a previous study neoplastic lymphocytes diagnostic of feline spinal lymphoma were Cytopathological and histopathological identified on CSF analysis in 35% of cats (4). Others (8) have studies reported that CSF was abnormal in 80% and diagnostic in 20% in a group of cats with NS lymphoma, pleocytosis associated with In the case of a non-diagnostic CSF, FNA or surgical biopsy increased TP was most commonly observed (65%), and similar of the lesion/s may be of greater diagnostic value than CSF alone to our study albuminocytological dissociation was present in (14). In the present study pathological studies were performed, 35%. In another study, the evaluation of CSF in 11 cats with either alone or combined, in samples obtained from the lesions lymphosarcoma revealed large neoplastic lymphocytes in 1 cat, by FNA, imprints, or tissue samples either by biopsy or necropsy. albuminocytological dissociation in 2 cats, and an increase In the 4 cases in which tests were combined, the same diagnosis in protein content and neutrophilic pleocytosis in 2 cats (2). was supported and further categorization in tumor types was Additionally, in cases of neurolymphomatosis CSF examination achieved in 2 of them. For the rest, a single test was able to might be useful in demonstrating a lymphocytic pleocytosis provide a diagnosis of lymphoma. (20, 28), in fact, in our group of 5 cats with lesions centered on In all the samples obtained by FNA, a high population the nerve roots CSF was retrieved in 3 cases and was diagnostic of morphologically abnormal lymphoid cells compatible with in 2 of them. lymphoma was observed. A further description was available In humans, CSF examination is included in the only in one case that was characterized as large granular recommendation of workup in patients with primary epidural lymphocyte lymphoma (and later confirmed by necropsy), lymphomas (11). A review of the systematic approach to the therefore, FNA of the lesions allowed the confirmation of diagnosis of suspected CNS lymphoma (13) concluded that at lymphoma but generally, typification was not achieved. FNA least one of the routine CSF indices is abnormal in more than can be obtained from the tumoral mass, more easily when a 80% of cases at the time of diagnosis. CSF cytology can provide paravertebral soft tissue mass or enlarged nerves are identified definitive diagnostic information in CNS lymphoma, and with by imaging, and/or from other masses of affected organs if the aid of immunohistochemical studies, it is possible to identify detected (35, 37). Small cell lymphomas are rare but can be Frontiers in Veterinary Science 19 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 suspected by FNA (14) and when necessary differentiation of 62.5% T-cell lymphomas and 37.5% B-cell lymphomas. from inflammatory lymphocytes can be supported by PARR. In Immunocytochemistry can be used to detect specific antigens humans, FNA is commonly used for NS lymphoma diagnosis. In on a cell’s surface which allows the differentiation of cell lines a study addressing paraspinal masses in 59 cases, FNA yielded a and of neoplasms subtypes, such as the B- and T-cell subtypes of definitive diagnosis in 66% of the cases and tissue biopsy yielded lymphoma (14). According to Harris et al. immunophenotyping no additional information. The sensitivity and specificity of FNA was helpful in some diagnoses of lymphoma, but not required for of paraspinal masses were 88 and 75%, respectively (38). many others. In general, in the absence of immunophenotyping, In surgical tissue biopsies, a lymphocytic proliferative those large cell lymphomas with basophilic cytoplasm tend to population of cells was present in all samples. When be of B-cell type, and those with water-clear cytoplasm tend to cytology from an intrasurgical lesion smear was performed be of T-cell type (16). The morphological and histopathologic it was compatible with lymphoma in 25% of cases, and distribution of nervous system lymphoma in cats has been further categorization was achieved in one case based on previously described (20, 23). The REAL classification of histopathological studies. Therefore, similarly to FNA samples, lymphoid neoplasms adopted by the WHO Harris, was a diagnosis of lymphoma was established by lesion smears adapted for animals and applied for the classification of feline but a further classification was not obtained. Surgical tissue lymphomas in an attempt to correlate histotypes and biological samples are good candidates for histological investigation (2) behavior (16, 23), this classification is based on both morphology and it has been reported that a preliminary diagnosis can be and immunophenotype, and on cytogenetics for some subtypes. facilitated by intraoperative smear cytology (20). Similarly This system requires immunophenotyping and does not provide in humans, the different tumor types are not discernable by a low/high grading system (16) and moreover, and as IHQ was neuroimaging, therefore histological evaluation of biopsied not available for all cats in the present study, the NCI WF system tissue is commonly necessary for the definitive diagnosis of was used to classify lymphoma subtypes based on histological neurotropic lymphoma. Macroscopically, extradural lymphoma studies as low, intermediate and high-grade. Most of the tumors typically appears as a focal soft, gray, non-encapsulated were categorized in the intermediate-grade (81.5%), followed by collection of neoplastic tissue in the epidural fat, that may the high-grade (14.8%) and only one (3.7%) was considered a invade the dura mater and spinal cord parenchyma and locally low-grade lymphoma. extend into the vertebral bodies or the underlying soft tissue The clinical significance of anatomical and (20). When intradural-extramedullary, lymphoma occurs as an cytomorphological evaluation in feline lymphomas has irregular mass within the subdural and subarachnoid spaces been indicated before in a study (40), which reported that and commonly infiltrating the spinal cord parenchyma or cats with globule leukocyte lymphoma showed significantly nerve roots (1, 20). Surgical nerve biopsies directed by imaging shorter survival than those with high-grade and other low-grade are also valuable in suspected cases of neurolymphomatosis lymphomas and cats with high-grade lymphomas showed (28). In humans, lymphoma affecting primarily peripheral significantly shorter survival than cats with other low-grade nerves is very rare, and diagnosis is based on MRI and surgical lymphomas. The incidence of different tumor types varies biopsy (37). greatly between studies. In one study (21), lymphomas were In postmortem anatomopathological studies of the case high-grade lymphoblastic or immunoblastic types in all cats. with an MRI localization of intramedullary multifocal, an In the study of Valli et al., about 30% of cat lymphomas were infiltrative leptomeningeal lymphoplasmacytic lymphomatosis of intermediate grade and about half were of high grade. was diagnosed, showing a more extensive infiltration than Only about 2% of cats had lymphoblastic lymphomas in the the imaging study, involving spinal roots and nerves. For lower categories of mitoses, and those classified as high-grade the rest of the studies, MRI depicted the extension of the lymphomas were more frequent than expected in the higher lesion infiltration in bone and soft tissues corresponding to categories of mitoses. In a study on non-neurological feline the pathological findings. Leptomeningeal lymphomatosis is lymphoma, there was no cytomorphological predominant characterized by widespread leptomeningeal infiltration by subtype (40). neoplastic lymphocytes (39) and has been reported in cats (20). Neoplastic lymphocytes invade the leptomeningeal space by direct extension from preexisting primary or metastatic CNS Outcome tumors or infiltrate through arachnoid vessels created by the tumor. Lymphomatous meningitis may be difficult to detect In our group of cats, 40.74% of them did not survive more grossly (20), and in this study was not detected by MRI. than 1 week after diagnosis. It has to be considered that data Necropsy allowed the confirmation of multicentric involvement from survival time since MRI longer than 1 week were only in a case that was suspected by MRI. available for 9/27 cats and ranged from 1 week to 13 months, but IHQ studies were performed in order to improve the as 5/9 cats were lost to follow up at that recorded time the actual characterization of the tumors in 8 cases with a result survival time for this group is unknown but may be greater. Frontiers in Veterinary Science 20 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 Survival time was recorded for 2 cases treated with In humans, diagnostic suspicion of primary CNS lymphoma chemotherapy alone and for 2 cases only surgically treated, on MRI may prompt early biopsy rather than conservative showing a wide variation (from 2 weeks to 13 months). management in the appropriate clinical setting (12). Moreover, Therefore, due to the low and dispersed numbers, a correlation the severity of necrosis on the histologic specimen on human between treatment and survival was considered difficult to primary brain lymphoma has been significatively correlated establish for those groups. For the 4 cases treated with surgery with hyperintense MR signal on T2W images and with rim combined with chemotherapy time of survival ranged from 1 to enhancement (12). Conversely, in our group of cats when 6 months (mean 2.75 months). A previous study (21) reported this correlation was studied, the presence of hyperintensity that a single cat treated with laminectomy and post-operative on T2W sequences decreased the probability of necrosis. chemotherapy had a prolonged remission (15.5 months). In our This finding may reflect a difference in the nature between group of cats that were similarly treated, the mean survival time human intracranial primary lymphoma and feline spinal cord was lower but it has to be taken into account that 2 cats were lost lymphomas, but it must be taken into account that in our study at 1 month and this may have lowered the actual survival time. In biopsy or necropsy were not available for all cats, therefore humans, when primary spinal epidural lymphoma is suspected the presence of tissue necrosis could have been missed in surgery is indicated to ameliorate the spinal cord compression some cases, thus making a bias in the correlation with the and to facilitate a specimen for histological diagnosis. Surgery MRI features. may be then followed by chemotherapy/radiotherapy (11). It has been reported that the prognosis for cats and dogs with brachial plexus lymphoma is unknown (41) Conclusion but one study of peripheral nerve sheath tumors in dogs reported that it is worsened when there is spinal cord In our group of 27 cats with spinal lymphoma, all of them involvement (42). In our study, cats were selected only if were evaluated for progressive neurological deficits, commonly the neurolymphomatosis invaded the spinal canal, therefore acute or subacute, with paraparesis as the most common a worse prognosis compared to an extracanal localization sign (62.9%), and pain present in 1/3 of the cases. Age at was expected. presentation appeared bimodal and related to FeLV status, When the cases were grouped in grades according to the with 40.7% of cats 2.5 years or younger (63.6% of them NCI WF (Table 6) the was no apparent correlation with the FeLV positive), and 44.4% of cats 8 years or older (16.7% survival time. In general, larger survival times corresponded to of them FeLV positive). In MRI studies spinal lymphomas intermediate grading, but there were exceptions (the only case were most commonly shown as a lumbar/lumbosacral epidural in the low grade died in the following days of MRI and one lesion, circumferential to the spinal cord, hyperintense on case with signs of necrosis on histopathological studies had the T2W images, and isointense on T1W images, with variable largest survival time with surgical treatment alone). However, a degrees of post-contrast enhancement and bone involvement. correlation was found between survival time and the presence of When the T2W signal was not hyperintense it increased the necrosis in histopathological studies, with lower survival times probability of the presence of necrosis in histopathological when necrosis was observed. studies as an unfavorable prognostic factor. The majority of The prognostic staging of feline lymphoma based on the lesions were focal (81.5%) but tended to extend to >1 lesion localization and distribution in I-V stages categorizes all vertebral body (66.6%). In the follow-up MRI studies performed paraspinal or epidural tumors regardless of other tumor sites in 4 cases new spinal lesions on different localizations and as stage III, and initial involvement of the CNS and/or bone extensions were present in 3 of them. Diagnostic confirmation of marrow in stage V (43). Although there is not a consensus spinal lymphoma was performed by CSF, FNA, surgical biopsy, about the clinical relevance of lymphoma staging there are necropsy, or a combination of them. For sample planning, indications that it affects prognosis (9). In cats, as in dogs the findings from the MRI studies regarding localization and and humans, the wide variation of lymphomas in tumor type distribution of the lesion/s were highly valuable. Antemortem and grade would translate to a wide variation in biologic diagnosis was achieved in 81.5% of cases. MRI allowed a behavior and treatment response (16). There are insufficient presumptive diagnosis and had a high concordance with studies on differences in the distribution of T, B, and NK cell pathological studies regarding lesion localization and extension. lymphomas, or low and high-grade lymphomas of the nervous The histopathology of the lesions was variable, with most of system (20). Lymphoid neoplasms show considerable variations the tumors graded as intermediate. Lymphoma with spinal in cellular origin and biological behavior. The present and involvement in the cat has a guarded prognosis. A fast and previous studies (40) illustrate the need for further studies accurate identification of the process, facilitated by MRI and on feline lymphomas that may facilitate classification into confirmed by cytology, may allow better counseling to the anatomical patterns and cytological subtypes. Categorization owners, with earlier and more precise treatments which may lead may be used to develop individual prognostic profiles (9). to prognosis improvement. Frontiers in Veterinary Science 21 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 Limitations Author contributions The main limitation of the present report is the retrospective VL: conception and design of the study, acquisition, design. Not all the data were available for all the cats and the analysis and interpretation of data, and drafting and actual survival times are not known for all the cats as some revising the article for intellectual content. JR, MB, and were lost to follow-up. The imaging studies were performed with MP: acquisition of data and drafting and revising the article different magnets and although the protocols were standardized for intellectual content. JM and MM: acquisition of data some differences were present. Histopathology was not available and revising the article for intellectual content. CB, TL, for all patients, and immunophenotyping allowing further and AG: acquisition of data. All authors approved the classification of specific tumor types was not performed in all submitted version. cases. Moreover, in the cases with a multifocal pattern, not all the lesions that were observed on MRI were sampled premortem, and the specimens may not fully reflect the histologic nature of Acknowledgments all lesions. Although this case series is the largest at present time, the We gratefully acknowledge Ester Blasco and Tamara Rivero relatively low number of cases, even including several referral for their technical assistance. We also thank Bruno Serra for his centers, is also a limitation and highlights the difficulty in assistance with statistical analysis. collecting well-documented cases. Conflict of interest Data availability statement The authors declare that the research was conducted in The raw data supporting the conclusions of this article will the absence of any commercial or financial relationships be made available by the authors, without undue reservation. that could be construed as a potential conflict of interest. Ethics statement Publisher’s note The animal study was reviewed and approved by authors declare that human ethics approval was not needed for this All claims expressed in this article are solely those study. Sampling and imaging procedures were held in a clinical of the authors and do not necessarily represent those environment as part of the diagnostic protocol for medical of their affiliated organizations, or those of the publisher, purposes, with the signed consent of the owners. The material the editors and the reviewers. Any product that may be used for the study did not require supplementary procedures evaluated in this article, or claim that may be made by or handling. Written informed consent was obtained from the its manufacturer, is not guaranteed or endorsed by the owners for the participation of their animals in this study. publisher. References 1. Marioni-Henry K, Vite CH, Newton AL, Van Winkle TJ. Prevalence 6. Palus V, Volk HA, Lamb CR, Targett MP, Cherubini GB. MRI features of diseases of the spinal cord of cats. J Vet Intern Med. (2004) 18:851–8. of CNS lymphoma in dogs and cats. Vet Radiol Ultrasound. (2012) 53:44–9. doi: 10.1111/j.1939-1676.2004.tb02632.x doi: 10.1111/j.1740-8261.2011.01872.x 2. Marioni-Henry K, Van Winkle TJ, Smith SH, Vite CH. Tumors affecting the 7. Auger M, Hecht S, Springer CM. Magnetic Resonance Imaging Features of spinal cord of cats: 85 cases (1980–2005). J Am Vet Med Assoc. (2008) 232:237–43. Extradural Spinal Neoplasia in 60 Dogs and Seven Cats. Front Vet Sci. (2021) doi: 10.2460/javma.232.2.237 7:610490. doi: 10.3389/fvets.2020.610490 3. Goncalves R, Platt SR, Llabres-Diaz FJ, Rogers KH, de Stefani A, Matiasek 8. Durand A, Keenihan E, Schweizer D, Maiolini A, Guevar J, Oevermann LA, et al. Clinical and magnetic resonance imaging findings in 92 cats with A, et al. Clinical and magnetic resonance imaging features of lymphoma clinical signs of spinal cord disease. J Feline Med Surg. (2009) 11:53– 9. involving the nervous system in cats. J Vet Intern Med. (2022) 36:679–93. doi: 10.1016/j.jfms.2008.05.001 doi: 10.1111/jvim.16350 4. Lane SB, Kornegay JN, Duncan JR, Oliver JE, Jr. Feline spinal 9. Moore A. Extranodal lymphoma in the cat: prognostic factors and treatment lymphosarcoma—a retrospective evaluation of 23 cats. J Vet Intern Med. options. J Feline Med Surg. (2013) 15:379–90. doi: 10.1177/1098612X13483236 (1994) 8:99–104. doi: 10.1111/j.1939-1676.1994.tb03205.x 10. Coulon A, Lafitte F, Hoang-Xuan K, Martin-Duverneuil N, Mokhtari 5. Bradshaw JM, Pearson GR, Gruffydd-Jones TJ. A retrospective study of 286 K, Blustajn J, et al. Radiographic findings in 37 cases of primary CNS cases of neurological disorders of the cat. J Comp Pathol. (2004) 131:112–20. lymphoma in immunocompetent patients. Eur Radiol. (2002) 12:329–40. doi: 10.1016/j.jcpa.2004.01.010 doi: 10.1007/s003300101037 Frontiers in Veterinary Science 22 frontiersin.org Lorenzo et al. 10.3389/fvets.2022.980414 11. Cugati G, Singh M, Pande A, Ramamurthi R, Balasubramanyam M, Sethi SK, 28. Hsueh CS, Tsai CY, Lee JC, Kao CL, Wang FI, Jeng CR, et al. et al. Primary spinal epidural lymphomas. J Craniovertebr Junction Spine. (2011) CD56+ B-cell Neurolymphomatosis in a Cat. J Comp Pathol. (2019) 169:25–9. 2:3–11. doi: 10.4103/0974-8237.85307 doi: 10.1016/j.jcpa.2019.03.004 12. Johnson BA, Fram EK, Johnson PC, Jacobowitz R. The variable MR 29. Beasley MJ, Hiebert EC, Daw DN, Alexander KJ, Gambino JM. appearance of primary lymphoma of the central nervous system: comparison with Neurolymphomatosis caused by T-cell lymphosarcoma in a cat: imaging histopathologic features. Am J Neuroradiol. (1997) 18:563–72. description and treatment review. JFMS Open Rep. (2019) 5:2055116919833534. doi: 10.1177/2055116919833534 13. Scott BJ, Douglas VC, Tihan T, Rubenstein JL, Josephson SA. A systematic approach to the diagnosis of suspected central nervous system lymphoma. JAMA 30. Elavarasi A, Dash D, Warrier AR, Bhatia R, Kumar L, Jain D, et al. Spinal Neurol. (2013) 70:311–9. doi: 10.1001/jamaneurol.2013.606 cord involvement in primary CNS lymphoma. J Clin Neurosci. (2018) 47:145–8. doi: 10.1016/j.jocn.2017.10.027 14. Twomey LN, Alleman AR. Cytodiagnosis of feline lymphoma. Comp Cont Educ Pract. (2005) 27:17–31. 31. Moulopoulos LA, Dimopoulos MA, Vourtsi A, Gouliamos A, Vlahos L. Bone lesions with soft-tissue mass: magnetic resonance imaging 15. Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, Vardiman diagnosis of lymphomatous involvement of the bone marrow versus J. Lymphoma classification–from controversy to consensus: the REAL and WHO multiple myeloma and bone metastases. Leuk Lymphoma. (1999) 34:179–84. Classification of lymphoid neoplasms. Ann Oncol. (2000) 11(Suppl. 1):3–10. doi: 10.3109/10428199909083395 doi: 10.1093/annonc/11.suppl_1.S3 32. Allett B, Hecht S. Magnetic resonance imaging findings in the spine of 16. Valli VE, Jacobs RM, Norris A, Couto CG, Morrison WB, McCaw D, et al. six dogs diagnosed with lymphoma. Vet Radiol Ultrasound. (2016) 57:154– 61. The histologic classification of 602 cases of feline lymphoproliferative disease using doi: 10.1111/vru.12340 the National Cancer Institute working formulation. J Vet Diagn Invest. (2000) 12:295–306. doi: 10.1177/104063870001200401 33. Wyatt S, De Risio L, Driver C, José-López R, Pivetta M, Beltran E. Neurological signs and MRI findings in 12 dogs with multiple myeloma. Vet Radiol 17. Higgins RJ, Bollen AW, Dickinson PJ, Sisó-Llonch S. Tumors of the nervous Ultrasound. (2019) 60:409–15. doi: 10.1111/vru.12759 system. In: Meuten DJ, editor. Tumors in Domestic Animals, 5th ed. IA: John Wiley and Sons (2017). p. 872–8. 34. Schmid MR, Hodler J, Vienne P, Binkert CA, Zanetti M. Bone marrow abnormalities of foot and ankle: STIR versus T1-weighted contrast- 18. Pi D, José-López R, Fernández F, Rabanal R, Mandara MT, Arús C, et al. enhanced fat-suppressed spin-echo MR imaging. Radiology. (2002) 224:463–9. Expression of FOXP3 in canine gliomas: immunohistochemical study of tumor- doi: 10.1148/radiol.2242011252 infiltrating regulatory lymphocytes. J Neuropathol Exp Neurol. (2020) 79:184–93. doi: 10.1093/jnen/nlz120 35. Singh M, Foster DJ, Child G, Lamb WA. Inflammatory cerebrospinal fluid analysis in cats: clinical diagnosis and outcome. J Feline Med Surg. (2005) 7:77–93. 19. Valli VE, Bienzle D, Meuten DJ 2017. Tumors of the hemolymphatic system. doi: 10.1016/j.jfms.2004.07.001 In: Meuten DJ, editor. Tumors in Domestic Animals. 5th ed. IA: John Wiley and Sons (2017). p. 203-321 doi: 10.1002/9781119181200.ch7 36. Feng L, Chen D, Zhou H, Shen C, Wang H, Sun X, et al. Spinal primary central nervous system lymphoma: case report and literature review. J Clin Neurosci. 20. Mandara MT, Motta L, Calò P. Distribution of feline lymphoma (2018) 50:16–9. doi: 10.1016/j.jocn.2018.01.034 in the central and peripheral nervous systems. Vet J. (2016) 216:109–16. doi: 10.1016/j.tvjl.2016.07.013 37. Gory G, Couturier J, Cauvin E, Fournel-Fleury C, Couturier L, Rault DN. Nervous system lymphoma with sciatic nerve involvement in two cats diagnosed 21. Spodnick GJ, Berg J, Moore FM, Cotter SM. Spinal lymphoma in cats–21 cases using computed tomography and ultrasound guided fine needle aspiration. Vlaams (1976-1989). J Am Vet Med Assoc. (1992) 200:373–6. Diergeneeskundig Tijdschrift. (2014) 83:107–12. doi: 10.21825/vdt.v83i3.16649 22. Rissi DR, McHale BJ, Miller AD. Primary nervous system lymphoma in cats. J 38. Wu JM, Sheth S, Ali SZ. Cytopathologic analysis of paraspinal masses: a study Vet Diagn Invest. (2022) 20:10406387221090281. doi: 10.1177/10406387221090281 of 59 cases with clinicoradiologic correlation. Diagn Cytopathol. (2005) 33:157–61. 23. Mello LS, Leite-Filho RV, Panziera W, Bandinelli MB, Sonne L, doi: 10.1002/dc.20319 Driemeier D, et al. Feline lymphoma in the nervous system: pathological, 39. Grossman SA, Krabak MJ. Leptomeningeal carcinomatosis. Cancer Treat Rev. immunohistochemical, and etiological aspects in 16 cats. Pesq Vet Bras. (2019) (1999) 25:103–19. doi: 10.1053/ctrv.1999.0119 39:393–401. doi: 10.1590/1678-5150-pvb-6295 24. Economu L, Stell A, O’Neill DG, Schofield I, Stevens K, Brodbelt D. Incidence 40. Sato H, Fujino Y, Chino J, Takahashi M, Fukushima K, Goto-Koshino Y, and risk factors for feline lymphoma in UK primary-care practice. J Small Anim et al. Prognostic analyses on anatomical and morphological classification of feline Pract. (2021) 62:97–106. doi: 10.1111/jsap.13266 lymphoma. J Vet Med Sci. (2014) 76:807–11. doi: 10.1292/jvms.13-0260 25. Troxel MT, Vite CH, Van Winkle TJ, Newton AL, Tiches D, Dayrell-Hart B, 41. Mellanby RJ, Jeffery ND, Baines EA, Woodger N, Herrtage ME. et al. Feline intracranial neoplasia: retrospective review of 160 cases (1985-2001). J Magnetic resonance imaging in the diagnosis of lymphoma involving Vet Intern Med. (2003) 17:850–9. doi: 10.1111/j.1939-1676.2003.tb02525.x the brachial plexus in a cat. Vet Radiol Ultrasound. (2003) 44:522–5. doi: 10.1111/j.1740-8261.2003.tb00500.x 26. Bray KY, Munana KR, Meichner K, White LA, Seiler G. Eosinophilic meningomyelitis associated with T-cell lymphoma in a cat. Vet Clin Pathol. (2016) 42. Brehm DM, Vite CH, Steinberg HS, Haviland J, van Winkle T. A retrospective 45:698–702. doi: 10.1111/vcp.12406 evaluation of 51 cases of peripheral nerve sheath tumors in the dog. J Am Anim Hosp Assoc. (1995) 31:349–59. doi: 10.5326/15473317-31-4-349 27. Meichner K, Kruse DB, Hirschberger J, Hartmann K. Changes in prevalence of progressive feline leukaemia virus infection in cats with lymphoma in Germany. 43. Mooney SC, Hayes AA. Lymphoma in the cat: an approach to diagnosis and Vet Rec. (2012) 171:348. doi: 10.1136/vr.100813 management. Semin Vet Med Surg Small Anim. (1986) 1:51–7. Frontiers in Veterinary Science 23 frontiersin.org